Engineering Fracture MecMnics, 1971,Vol. 3, pp. 435-451. Pergamon Press.

Printed in Great Britain

PLASTIC INTENSITY FACTORS FOR

CRACKED PLATESt

PETER D. HILTON and JOHN W. HUTCHINSON

Harvard University, Cambridge, Mass. 02138, U.S.A.

Abstract-An elastic-plastic analysis is performed for two problems relevant to fracture mechanics: a semi
infinite body with an edge crack in a far out-of-plane shearing field and an infinite plate under plane stress
conditions containing a finite line crack in a remote tensile field. Amplitudes of the dominant singularity in
the plastic region at the crack tip, the plastic stress and strain intensity factors, are calculated for applied
stress levels approaching the yield stress. A technique is developed for using the dominant singular solution
in conjunction with the finite element method to make accurate calculations for the near-tip fields. Addition
ally, a comparative study of deformation theory with flow theory is performed for cracks in an anti-plane
shear field. Elastic fracture mechanics is extended to high levels of applied stress for which the plastic zone
is no longer small compared to the crack length by relating the critical stress for fracture initiation to the
plastic intensity factors.

INTRODUCTION
AN ELASTiC-plastic analysis of plates of hardening material containing cracks will be
carried out. This analysis is then employed to extend classical fracture mechanics to
the case when the plastic zone about the crack tip can no longer be considered small.
Work pertinent to this area of fracture mechanics has been reviewed by McClintock
and Irwinj l] and very recently by Rice [2].
Initiation of crack growth depends on the stress and strain fields in the immediate
vicinity of the crack tip. For the cases considered in the present study, cracks in
remote anti-plane shear and tensile fields, these near-tip solutions are known except
for their amplitudes, the plastic stress and strain intensity factors which relate the be
havior at the crack tip to the geometry and applied stress. In the small scale yielding
range, when the plastic zone about the crack tip is small in comparison with the crack
length, the plastic intensity factors can be related directly to the elastic stress intensity
factor. Results of this type, which will be further discussed, are due to Hult and
McClintock[3], Neuber[4] and Rice[5] for the anti-plane shear case and to Rice and
Rosengren[6] and Hutchinson [7, 8] for the tensile case. At higher values of applied
stress, in what will be referred to as the large scale yielding range, the plastic zone is
no longer small compared to the crack length and the elastic stress intensity factor is
no longer relevant without some modification.
Use of plastic intensity factors is in no way restricted by the extent of the plastic
zone; and thus the concept of a critical value of the plastic stress or strain intensity
factor can be introduced and employed in much the same way as the elastic stress
intensity factor is used in classical fracture mechanics. A numerical procedure is
developed here which combines details of the dominant singularity in the plastic region
with the finite element technique for accurate computation of the plastic intensity
factors. Standard numerical techniques which have been applied to these crack prob
lems with no special treatment for the singularity at the crack tip fail to give accurate
solutions about the tip, the region of prime interest.
[Presented at the Third National Symposium on Fracture Mechanics, Lehigh University, Bethlehem, Pa.,
August 25-27, 1969.
435

436 P. D. HILTO N and J . W. HU TC H INSON

T he probl em of a se mi-infinite bod y with an edge crac k s ubjected to a far o ut-of

plane shear field is use d to illustra te the techniqu es develope d. Accuracy esti ma tes fo r
thi s pro blem fo r th e case of a deformati on theo ry of pla sticity are es tablished throu gh
co mparisons with the re sults which Rice [5] o btained with a differen t techniqu e. F ur
th er , th e ca lc ula tions are re peated with an inc re me nta l th eo ry to stud y th e di screp ancies
be tween the p re dictio ns of deformatio n theory and those bas ed on flow theory . Th e
major co ntri butio n of thi s wo rk is the result s w hich have bee n obta ine d fo r a finit e line
c rack in an infinit e sheet und er plan e stress co nditions subjec ted to a te nsile field .

PLASTI C STRESS AND STRA IN INT ENSITY FAC T O RS

T he do minant singu lar term of the elastic so lutio n in the ne ighborhood of a crack
tip is wr itten as

(1)

where ( I'. 0) are pol ar coordinates ce nte re d at the tip. f A co llection of e las tic so lutio ns
to crack proble ms for a va riety of geo metries an d load ing co nd itio ns is given in the
review paper by P aris a nd Sih [9] . Fo r all loa dings and geomet ries suc h th at th e stress
fields are symmetric (or ant isyrnrnetric), th e circumfe ren tial var iat ion of the st ress field,
&;;(8), is independen t of additional features of both the geome tr y a nd the bou ndary
co nditions of the pa rticular prob le m co nsidered . O nly the a mp litude K ef' co mmon ly
ca lled the st ress inte ns ity fac to r, varies fro m problem to pro blem. Fract ure initiat ion
takes place , for a given set of co nd itions. whe n the amplitude of th e st res s field at th e
crack tip re ach e s a critica l va lue . Th is approac h pre sum es th at the plastic l o ne is s uffi
cie ntly s ma ll so t hat ( I) rep resents th e stre ss field accura te ly in the ne ighbo rhood of th e
crack tip outs ide th e p lastic l one.
It is no w st andard proced nr e to expe rimentally de termine t he va lue of the stress
intensity factor at which fracture initiates an d to apply the res ult s to initial frac t ure
predict ion fo r othe r co nfiguratio ns. T here are a nu mbe r of necessaril y re st ricti ve condi
tion s w hic h must be p lace d on this technique , and det ails such as co nditio ns ap proac h
ing plane stress and p lane strain mus t be different iate d as di scussed in [10 ].
A simp le examp le wh ich illustr ate s elas tic frac ture mec ha nics is that of a n infinite
(T u = ~ a'l j {H) (1 )
v 2m"
whe re ( I'. 0) are po lar coordinates ce nte red at th e tip . f A co llect ion of e las tic so lutions
to c rack proble ms for a va riety of geometries a nd loa ding co nd itio ns is give n in the
rev iew paper by P aris and Sih [9]. Fo r all loadings and geomet ries such th at th e stress
fields are symme tr ic (or a nt isy rnmet ric) , th e circu mfe re ntial va riation of th e stress field ,
lfiJ(8 ) , is independent of additional feat ures of both the geometry a nd the bou ndary
co nditio ns of the parti cul a r pro blem co nsidered . On ly the amp litude K ef, co mmon ly
ca lled th e stress in tens ity facto r, varies fro m problem to problem. Fractu re initi at ion
takes place, fo r a given set of cond ition s . when the amplitude of th e st res s field at the
crack tip reach e s a critica l va lue. Thi s approac h pre sum es th a i the plasti c l o ne is s uffi
cie ntly s ma ll so t hat ( I ) repr esents the st re ss field accu rate ly in the ne ighb o rhood of th e
crack tip outs ide the p lastic l one.
It is no w st andard proced nr e to expe rimentall y de termine t he va lue of the st re ss
inte nsity factor at which fracture initi ates and to apply the res ult s to initial fract ure
predict ion for ot her configurat io ns . There are a numbe r of necessaril y re stricti ve con di
tio ns w hic h mu s t be placed o n this tech niq ue , and details suc h as co nd itio ns a pproach
ing plane stress and p lane strai n must be different iated a s discussed in [10 ].
A simp le examp le w hich illustr ate s elast ic fracture mech an ics is that of a n infin ite
 Plastic intensity factors 437

As discussed irr the Introduction, the dominant singularity in the plastic zone for
hardening materials will be exploited to extend elastic fracture mechanics, and the
plastic stress (or strain) intensity factor associated with the dominant term for the near
tip fields will be used in a manner analogous to the elastic stress intensity factor to
correlate fracture initiation results.
A small strain formulation of plasticity is used as the basis for the present studies.
The tensile stress-strain relations

<T~1
(3)
<T>1

are chosen to model the tensile elastic-plastic behavior of the material where the
material coefficient is n.
For the present discussion a total deformation theory of plasticity is employed.
Plastic deformation is assumed to be independent of the hydrostatic component of
stress, <Tkk, and completely determined by the first invariant of the stress deviator Sij =
<Tlj -!<TkkOij. This invariant, the 'effective stress' <Te , is defined by <Tl = fSijsij. For
simple tension <Te = <T and the Mises yield condition is <Te 1. The generalized stress
strain relationship which reduces to (3) for simple tension is

ejj/(l + v);
Sij =
[ee/(l + v)](l-nlln eij/(l + v); <Te > 1

where v is Poisson's ratio, the strain deviator is elj = ij-1ppOij and the effective strain
ee is defined by ee2 ~BiAj. In the subsequent formulation it will be assumed that no
unloading occurs for the monotonic loading histories considered in this paper. This
necessitates an a posteriori check of the solution to assure that this assumption has not
been violated.
As in the elastic analysis, the dominant singular term of the asymptotic expansion
at the crack tip can be written as an intensity factor multiplying a function which is
independent of geometry and boundary conditions. For plane stress, plane strain, and
anti-plane shear the dominant fields associated with a 'power hardening material' can
be written in the followingform

(5)

where K (Ka)n and where the dimensionless functions of 0, a-ij and Eij, are detailed
for plane stress and plane strain in [8].t
In the region at the crack tip dominated by the singularity solution (5) the deforma
tion is proportional, that is, the relative magnitudes of the stress components do not
change with increasing applied load. For this reason, (5) is valid for flow theory as well
as deformation theory; however, in general, the amplitudes will differ depending on
which theory is employed.

tUnlike in [8], here r is the distance from the crack tip and has not been normalized by the crack length 2a.
438 P. D. HILTON and J. W. HUTCHINSON
p{a.si"t' c
It must be emphasized that the concept ofa eIa!;tic strain
intensity factor associ
ated with a dominant singularity is tied to hardening materials. The distribution of
strain at the tip of a crack in a perfectly plastic material is no longer in one to one
correspondence with the stress distribution and, in most cases, iPJ) will depend on
the applied stress. Alternate measures of the crack tip deformation are possible in this
limiting case, but they will not be dwelt on here.
Asymptotic expressions relating the plastic intensity factors to the elastic stress
intensity factor, valid for small scale yielding, can be obtained using the methods out
lined in [6, 7]. These results are of the form

(6)

and values of the coefficients en, which depend on the hardening coefficient, are given
for a number of cases for both plane stress and plane strain in [8]. For an infinite plate
containing a line crack of length 2 a, K(T and K E can be rewritten as

(7)

In the large scale yielding range (7) no longer holds and in its place are expressions of
the form

(8)

where/will be calculated up to values of (T'" = 09 for the problems investigated in this

paper.
In direct analog to elastic fracture mechanics, critical values for the plastic stress
and strain intensity factors associated with fracture initiation, K(Tc and K/ = (K(Tc)n,
are introduced. In the small scale yielding range, when K(T' K E and K e1 are directly re
lated by (6), the predictions based on the plastic intensity factors are identical to those
based on the elastic stress intensity factor. When large scale yielding occurs, K(T and
K el are no longer directly related, independent of the crack length, and it will be
necessary to use (8) to modify the elastic predictions.

ELASTIC-PLASTIC ANALYSIS OF CRACKED BODIES IN

ANTI-PLANE SHEAR

The equations governing the anti-plane shear case are simpler in form than those for
in-plane problems so that it is convenient to use this problem as a model to exhibit the
techniques developed in the present work. Further, the numerical results can be com
pared to those that Rice [5] obtained with a different technique.
Consider an infinite body containing a finite straight crack subjected to a remote,
out-of-plane shearing field. From symmetry considerations, this is the same problem
as that of a semi-infinite body containing an edge crack (Fig. 1). Let w represent the
displacement normal to the plane of the plate. The only non-zero strains are 'Yxz W. x
and 'YIIZ = w,lI .t The dominant singular solution for a power hardening material of the
form given by (4) has been derived by Neuber[4] and Rice [5]. With polar coordinates

tin this section only. (T".. Tv" K,,) ('T",,, Tw K,,)!Ty where 'Ty is the yield stress in shear and (".r.' "y K.) =
(1.r., yye- K.)/y y where yy ""TylG and w = wl"iY
 Plastic intensity factors 439

Fig. l(a). A semi-infinite body with an edge crack in a remote anti-plane shear field.

I -CD 2
-(!Ty )
Ro' 2L

n-I) Ro
130:(n+r

Fig. 1(b). The elastic-plastic boundary for small scale yielding in anti-plane shear.

(r, (J) centered at the crack tip, this solution can be reduced to the form

{T:rZ} = K(T [h(J)]lIln+ll {-Sin p }

T II Z r cosp

{'Y:rz}
'YYZ
= K [tll!!l]nlln+lJ
r
{-Sin
cos p
p}
(9)
h(J )]-IJ(n+ll .
w=K [-r- Slnp

where

For small scale yielding, the character of the entire field is exceptionally simple as
has been discussed in [5]. The plastic zone bounded by a circle of radius T"'"a/2 with its
center shifted a distance T""a/2(n--l/n+ 1) ahead of the crack tip is shown in Fig. l(b).
The plastic stress intensity factor is given by
- ",2/(n+ll a l/(n+ll .
K (T-T (10)

Moreover, (9) is the full solution everywhere in the plastic zone. The plastic deforma
tion is exactly proportional and this solution is also a solution to J 2 flow theory.
440 P. D. HILTON and J. W. HUTCHINSON

Attention is now directed to the large scale yielding problem for which the shape of
the plastic zone becomes non-circular and (10) for K; no longer holds. A numerical
procedure is developed to connect, or match, the field at the crack tip which is governed
by the dominant singularity to the uniform stress field (Tz z = 0, T IIZ = TOO) far from the
crack. A finite element technique is specialized to this purpose. The variational prin
ciple of minimum potential energy for deformation theory plasticity forms the basis for
this method. A modified potential energy functional is introduced next which remains
finite for all admissible displacement fields and can, therefore, be applied to the infinite
domain under consideration.
Let r R be a circular arc centered at the intersection of the axes and of sufficient ra
dius, R, so that the plastic region is contained within it as shown in Fig. 2. Designate the

Fig. 2. The division of the quarter plane into two regions associated with the modified potential
energy functional.

region outside r R as region A 2 , the region inside r R as A I' In A 2 , the displacements are
written for the general case as the sum of ul and Ui' where UfO is chosen in such a way
that Ui = (u; Uj0) is of order (1/ r) for large R.
The modified potential energy functional is

MPE = JSED(ut)dA + JSED(ai)dA

Al A2
(11)

where SED (u;) = f

Eli

(J"iJdEij, SED(u;) = !UijE,j, nj is the outward unit normal to r R and

(J"~ are the stresses derived from Uto. It can be shown that the functional MPE for a
monotonically hardening material is minimized by the exact solution among all admiss
ible fields U; in A 1 and 17, in the infinite domain A 2 In the present application, WO = T"'y
which corresponds to the uniform stress state approached far from the crack, and
(J"~ njat TOOW sin () on I'R'
In A 2 , the solution is expressible in terms of an analytic function of the complex
variable z, cP(z), where

w = Real[cP(z)] and Tzz-iT y Z = cP'(z).

 Plastic intensity factors 441

The symmetry conditions are used to write the Laurent series for <p(z) about the origin
as
"" ia z-2m+l
>(z)=-iT""z+ ~ <Po+<f) for Izl;;:o R (12)
m
m=l

where <Po = -iT""z and the am are real. The series is convergent on and outside of f R
oe r;
R ""
w(R,8)=-t/r""sin8+ ~ amR-2m+l sin(2m-l)8. (13)
m=l

The coefficients am are related to the displacement field along I'R by

1f12
am = -R 21'''''131 m +~ R2m-1 f w(R, 8) sin (2m -1)8 d8 (14)
o

where 131m 1 for m = 1 and 131m =cO for m - 1. In what follows, symmetry is ex
ploited and integrations are performed over only the first quadrant of the x, y plane. The
contribution fromA 2 to MPE is
"" 1f/2
f SED(w)dA ! f rdr f d8(<f}'~')=i ~ am2(2m-l)R-4m+2.
A. R 0 m=l

Attention is now focused on the immediate vicinity of the crack tip. Consider a
circular are, f 1 , centered at the tip and contained in the fully plastic region. If ro, the
radius to f 1 , is taken sufficiently small, then the dominant term of the expansion (9),
which is asymptotically correct at the crack tip, is a good representation to the full
solution on and within [1' The strain energy of the region within [1' calculated from the
dominant singular solution is
Til TT' 71'

n_!:if si? 2p d8
SE = _n_f rdrfd8(r n+ 1) = K <n+llIn_ KE<n+Oln Sn
n+ 1 E n+ 12 sin 8
o 0 0

and the displacement field on [1 is given by

The finalform ofthe modifiedpotential energy functional can be written as

MPE=KE<n+O/nSn+ f SED(w)dA+i-~am2(2m 1)R-4m+2 - f 1''''' sin 8 w(R, 8)Rd8

Ai 1 rR
(15)
where Ai is the area between I', and f R The displacement field along [1 is now known
except for its amplitude K E which will be determined through the minimization of
MPE.
442 P. D. HILTON and J. W. HUTCHINSON

A finite element technique is employed to provide a representation of the solution

in the region between [1 and I'R' This region is divided into a triangular grid pattern.
Over each element, the displacement field is approximated by a general linear function;
thus the strain components in each element are constant[13]. The strain energy of an
element is expressed in terms of the displacements at the three nodes of the triangular
element. For a typical element, the strain energy is
3
SE = D I
j,j=1
BUWjWj

where Bij is a symmetric matrix whose elements depend only on the triangle geometry
and D is a nonlinear coefficient related to the element stiffness and given by D 1 if
the element is elastic and, if not, by

The Laurent series for the solution in A 2 is terminated at a finite number of terms
arbitrarily chosen to equal the number of nodal points along I'R' The coefficients am are
calculated in terms of the displacements at the nodes along f R using (14).
The above procedure results in a discretization of the modified potential energy
functional which depends on the plastic strain intensity factor K. and the nodal dis
placements Wj' Equilibrium equations associated with the minimization of MPE with
respect to each of these parameters are given by

oMPE(K., WI)
OWj
o j= I,M (16)

where M is the total number of nodal points in the grid pattern. Equations (16) are
nonlinear algebraic in form and an iterative procedure is used to solve them. A choice
of initial values for the coefficients D of the stiffness matrix for each element and for
the stiffness parameter, K.o-nnn, for the inner core region bounded by f 1 renders the
system of equations linear in the nodal deflections and K . This large system of equa
tions is solved and the coefficients D and K.o-n>/n are recalculated. The iterative
process is continued until convergence is attained.
The accuracy of these calculations depends on the number and distribution of
elements and on the radius ro chosen for the inner boundary. A grid pattern consisting
of 546 elements with the inside radius ro equal to two percent of the crack length was
chosen. Eleven node points are taken along the half-circular arcs near the crack tip.
The resulting procedure for a given value of TOO converged within six iterations for all
cases considered with a reasonable set of initial guesses. The strain invariant for each
element increased with increasing values of the loading parameter. Thus unloading did
not occur and its exclusion in the formation of the stress-strain relations is justified. A
comparison with the results by Rice [5J indicated that the accuracy ofthe technique for
a material with high strain hardening capacity was quite good and the results were
never more than two percent in error for n = 10/3.However, when the strain hardening
capacity diminished as in the case of n = 10, values for the plastic strain intensity factor
were as much as five percent low for large values of the applied stress.
The deviation of the plastic strain intensity factor from the small scale yielding value
as given by (10) is shown in Fig. 3(a). Specifically, the ratio A(T oo) K./(K.)s.s.II., which
 Plastic intensity factors 443

1.4
1.3
1.2
K. 11
(K.)u.y. I.~I- -

.9

Fig. 3(a). The plastic strain intensity factor as a function of the applied stress for anti-plane
shear.

Fig. 3(b). The critical initiation stress for large scale yielding, anti-plane shear.

by (7) and (8) is independent of the crack length, is plotted against T'" == "f"'Fiy These
results are now employed in conjunction with the fracture criterion, that K. reach the
critical value K/ when the crack starts to extend, to predict the critical initiation stress
in the large scale yielding range. Using the critical elastic stress intensity factor and
the identity (6) in the small scale yielding range, an implicit equation for Tj'" results:

7:'" = K%z 1 (17)

j [.\(Tt')] n+112n "\.1;;;'
In this form it is clear by comparison with (2) that K;Z/[.\(Tt')] n+ 1/2n can be regarded as
the modified 'elastic' stress intensity factor. Plots of lft,/lfy vs. K;z/'lf y y:;;;;, are given in
Fig. 3(b); and thus, deviations from the 45 line at roughly half the yield stress indicate
a departure from the predictions of elastic fracture mechanics.
The plastic intensity factors for the anti-plane shear problem treated above were
recalculated using an incremental theory of plasticity. The same uniaxial stress-strain
curve was used in conjunction with J 2 flow theory. As has already been noted, the
form of the dominant singularity given by (5) and (9) will still hold for this incremental
 444 P. D. HILTON and J. W. HUTCHINSON

theory but the amplitude out of the small scale yielding range will, in general, differ
from the predictions of deformation theory.
Most of the aspects of the incremental theory calculation are similar to the method
discussed in the foregoing. A modified strain rate functional, specialized to the incre
mental problem, is again used together with the embedded singularity. In this calcula
tion, however, the load is increased in small steps and a sequence of linear incremental
problems is solved. Now, increments of the intensity factor K., as well as the nodal
displacements Wi, are solved. Since the small scale yielding solution is correct for J 2
flow theory as well as deformation theory, it is used to 'start' the incremental solution
at a sufficiently low value of the applied stress. Accuracy checks were made in several
ways. First, the effect of varying the load increments was determined, and second, this
same incremental procedure was applied to the deformation theory problem for which
a comparison with known results was possible. The plastic strain intensity factors cal
culated on this basis for flow theory for the cases n = 10/3 and n = 10 up to the value
7 = 09 were only very slightly below the corresponding predictions of deformation
00

theory and were within the accuracy of the numerical method.

To further demonstrate the versatility of the techniques developed in this section,
the linear-elasticity problem was also solved with the same grid pattern. The calcu
lated value for the stress intensity factor was well within one percent of the known
value. Thus, the method described for combining the knowledge of the character of the
singular solution with a finite element technique may tum out to be a useful technique
for the determination of elastic intensity factors associated with complicated geometries.
Other work along these lines is already underway[II], including a fairly sophisticated
analysis of this type of procedure [12].

CRACKED PLATES IN TENSILE FIELDS

The techniques which have been developed for the anti-plane shear problem are
applicable to the tensile case with only minor revision. In this section both the small
and large scale yielding problem for an infinite plate under plane stress conditions con
taining a finite line crack in a far tensile field will be treated. Deformation theory is
employed.
The extended Michell theorem[14] is exploited to permit consideration of an in
compressible material. This considerably simplifies the stress-strain relations in the
plastic region. The theorem states that for problems with tractions specified on all
boundaries, the stress field is independent of Poisson's ratio. Note that the features of
interest, namely, the plastic stress intensity factor and the plastic strain intensity fac
tor, which is related to it by K. = (Ku)n, as well as the elastic-plastic boundary, are
independent of Poisson's ratio for plane stress conditions.

Small scale yielding analysis

Here our primary interests are to obtain an accurate description of the elastic-plastic
boundary and to determine the accuracy that this type of calculation, which will later
be employed for the general problem, gives for the plastic intensity factor. The small
scale yielding problem is the asymptotic problem posed for the case in which the plastic
zone is very small in comparison with the crack length, 2a, so that it is embedded in the
elastic field dominated by the singular elastic solution. Mathematically, the problem
treated is a plate with a semi-infinite crack in a far stress field which is the elastic

I
 Plastic intensity factors 445

singular solution, i.e.

5/4 cos 8/2 1/4 cos 38/2

;
= .;;:; 1/4 sin 8/2 + 1/4 sin 38/2 {I 8)
v27Tr

3/4 cos 8/2 + 1/4cos 38/2

where e; ifoov:;;,;.
The solution asymptotically close to the crack tip is the dominant singular solution
whose form is given in (5). The circumferential variation of the fields (iTl,f(8) and EjJ(8
associated with the singular solution are reproduced from Ref. [8] in Fig. 4. For the

15r--------------,

0-3
o 7'/2
8

Fig. 4.IJ-Variationofstresses and strains at the tip ofa tensile crack for plane stress (from (7)).

small scale yielding problem, the plastic strain intensity factor is given by (7). The
values for Cn given in [8] which will be used here for comparison purposes are

C = {O'949 n= 3 {I 9)
n 1.004 n = 9.

The following non-dimensionalization collapses the set of solutions corresponding

to different values of (J'-OO into one set of similar solutions:
446 P. D. HILTON and J. W. HUTCHINSON

"V ) = 1 (Cf
(u, a y
(f'"
)2(Ey'
U
Ey ,
ii) (l,y) = ~(:~r (x,y)
" = ~
Et; (20)
Ey

where u, x, a'ij, and EU satisfy the same equations as ii, x, (Til' and Eil respectively. The
far stress field with this non-dimensionalization is the same (18) with K e,/V'Ziii re
placed by 1/"I/iJ. The non-dimensionalized plastic intensity factors are KIT = Cn and
KE = (cn)n. This non-dimensionalization is only for the small scale yielding problem and
will be used exclusively throughout this section. For convenience the "" symbol will
be omitted henceforth.
As in the anti-plane shear case, the plane is divided into two regions associated with
the modified potential energy functional. Denote as I'Rs a circular arc centered at the
crack tip and of sufficient radius R, so that the plastic zone is contained within it. In
region A 2 , outside f R s , the solution is written in terms of the Muskhelishvili[15] func
tions ~ and e where (with lJ = 1)

a'1I-a';r + 2ia'ZlI = 2(Z~" + .p')

a';r + a'1I = 2(~' +1>') (21)
u + iv = 5/2~ - 3/2(z4}' + $).

The expansions for ~ and .p which satisfy the stress-free condition along the crack and
symmetry are

~ = zl/2/V2 + L'" llmZ-m+112 = ~o + (f,

m=l (22)
.p = Z1l2/2V2 + ~ (m + 1/2)amCm+112 = .po + .p
m=l
where each am is real, and ~o = ZlI2/V2 and .po = ZlI2/2V2 are the terms which corres
pond to the far field of the dominant singularity of the elastic solution. The unknown
coefficients am can be expressed in terms of the displacements along I'R s , u(R, 8), etc.
as
'tI'

am =Rm-l/2-.l.j[(U+iV)
s 5," .
(uo+ivo)]ei(m-l/2J/ld8.
e

The contribution from a, to the modified potential energy functional is found to be

SE2 ( il ) = I m=l
L'" (-m+ 1/2)[(1+3(m+ 1/2)2)am 2
r - 2m+1

-3(m+ 3/2)(-m +5/2)amam+2r2m-l].

A circular arc I', of radius ro is centered at the crack tip and the dominant singular
solution described previously is used to obtain the fields on and within fl' The values
 Plastic intensity factors 44 7

for the circumferential variation of the displacements along [1 were obtained from the
calculations made in [7]. This information is used, just as in the anti-plane shear case ,
to determine the contribution of the inner core to the modified potential energy.
The region between [1 and [R s is divided into a node pattern by a series of concent
ric circul ar arcs and a set of radial lines. The resulting quadrangles are each divided
into two triangles , and in this way the finite element technique with constant strain,
triangular elements is used to connect the region between [1 and [ R' Thi s leads to a
set of nonlinear algebraic equations for K. and the nodal displacements who se solution
is carried out in the same way as it was for the anti-plane shear problem.
Computations were performed for two values of the hardening coefficient , n = 3 and
9. In each case the inside radius Po was taken to be 003 and the grid pattern consisted
of 1080 elements with 21 nodes on each half circular arc . For n = 3, this resulted in a
numerical value for the plastic strain intensity factor which was 3 per cent above the
known value (19). When the material hardening capacity was decreased to n = 9, the
calculated value was 2 per cent higher . Plots of the elastic-pl astic bound aries in non
dimensional similarity coordinates (x, 5') are pre sented in Fig. 5. A lso included is a

I
_ l'OT
1\ , .1...12)2
Y L \ 5 (I) I
75 T

.25 .5 75 10 1 25
1\ :
x
z.L (!!:L)2
?y O>

calculated val:~ <JvJa~1 2<t~e?l ~c~~thlg'h~e~ < p1oy; o~rihle '~t~.iitic~pl~ti2 "bRot~d aries in non
dimensional similarity coordinates (x, 5') are pre sented in Fig. 5. A lso included is a

I
_ l'OT
1\ , .1...12)2
Y L \ 5(I) I
75T

.25 .5 75 10 125
f'o
x
: z.L (!!:L)2
?y O>

Fig. 5. Elastic-plastic bound aries for small scale vieldinz in a ten sile field.
 448 P. D. HILTON and J. W. HUTCHINSON

sions for <b and tjI which satisfy the symmetry conditions about both axes for Iz I ; ; . R,
are
(f> = ~'" ~m_lZ-2m+1
m=l
00
~
~
b2m-l z-2m+l
m=l

where am and bm are real. Outside r R, <P = <Po + <P and .p .po + tjI. The coefficients am
and bm are related to the displacements along r R by

m= 1,3,5 ...

R2rr'"
- - -4R
6
-
31T
f
1T/2

(u+iv)e- 18 dO (23)
o
1T/2

bm=R2(m-2)am_2-4~m f (u+iv)e- i m 6 dO m 3,5,7...

o
The procedure from here follows that for the anti-plane shear problem. The con
tribution from A 2 to the modified potential energy functional is written in terms of the
coefficients am and bm which are in tum related to the displacement field along r R
through (23). The dominant singular solution discussed for small scale yielding is used
to describe the solution on and within the small circular arc fl' Its amplitude, K E , is to
be found through the minimization of the modified potential energy functional. The
region between f l and f R is divided into a triangular grid pattern and the finite element
technique is employed to connect the singular solution to the far field.
The computations for the large scale yielding problem were performed with a grid
consisting of 1066 elements and an inside radius To of two percent of the half crack
length. The calculated values for the plastic strain intensity factor in the large scale
yielding range associated with this grid are expected to be low by several per cent paral
leling the findings for anti-plane shear. A check was made on the assumed condition
of loading inherent in this formulation. Careful examination of the solution over the
range of the loading parameter investigated here indicated that with monotonically in
creasing values of the load parameter, the stress invariant everywhere in the field also
increased monotonically.
The results are presented in the same form as they were for the anti-plane shear
problem. In the large scale yielding range, the plastic strain intensity factor is KE =
[f(rr W anJ( n + ll , while the formula for the plastic strain intensity factor for the small
OO

scale yielding problem is given by (7). In Fig. 6(a). A(rr KE/(KE) s.s. s, is plotted
OO
(

against a" "" iJoo/Uy. The modifiedformula for the crack initiation stress is

and plots of rrloo/rry vs. K~l/rry v;a are given in Fig. 6(b) for n = 3 and 9.
The elastic-plastic boundaries corresponding to the above calculations are exhibited
in Fig. 7. At an applied stress level of half the yield stress the maximum extent of the

i
 Plastic intensity factors 449

1.3
1.2
J5..s.. 1.1
(K.)s.s.y 1.01-----.-=:::.::::;-

.9

Fig. 6(a). The plastic strain intensity factor as a function of the applied stress for tensile loading.

io

.8

.4

.2

-c
~ _1_
'i'y .fTi'O
Fig. 6(b). The critical initiation stress for large scale yielding, tension.

elastic-plastic boundary is already greater than 40 per cent of the half crack length.
Even so, elastic fracture mechanics, which is based on the dominant singularity of the
purely elastic solution, still gives quite accurate predictions for the critical initiation
stress at this level of applied stress.
The present predictions are now compared with predictions for fracture initiation
based on two perfect-plasticity models. The first, due to Irwint, is a semi-empirical
adjustment of the elastic stress intensity factor to take into account the size of the
plastic zone. Irwin makes use of the plastic zone size predicted for the perfect-plasti
city, anti-plane shear problem and argues that the effective crack length is increased
by an amount comparable to the size of the plastic zone. For the present problem,
Irwin's suggestion for the modified elastic stress intensity factor is

The predictions for the critical stress based on this are included in Fig. 6(b). Irwin's
formula is quite good for low hardening materials over the range of the loading para
meter plotted.
f Private communication, see also [1].
450 P. D. HILTON and J. W. HUTCHINSON

2.4

2.2

2.0

1.8

1.6

Y 1.4
L
1.2

1.0

.8

.6

.4

.2

0
2.0 2.2 2.4 2.6
liIL

Fig. 7. Elastic-plastic boundaries for a crack subjected to a remote field.The half crack length is
unity and the crack tip is taken at s = O.

The Dugdale-Barenblatt model, for the perfect-plasticity treatment of necking

ahead of the crack tip in thin sheets of certain metals, is the basis of the second com
parison. The crack opening displacement, the separation which occurs between the top
and bottom faces of the crack at the tip, as predicted by this model[2] is

0= ~ y log [sec (0''''' I)]' (24)

The associated fracture initiation criterion for a sequence of sheets with identical
thicknesses is that the crack opening displacement reach a critical value oC. For small
scale yielding this criterion leads to

Where there is no restriction on the applied stress, the critical stress is given by (24)
with 0 identified with OC. The resulting predictions fall between the present results for
n = 9 and those from Irwin's formula.
Rice [2, p. 293] has discussed the extension of fracture mechanics into the large
scale yielding range using results from the anti-plane shear problem as well as from the
 Plastic intensity factors 451

Dugdale-Barenblatt modeL Instead of emphasizing the plastic intensity factors, he

uses the value of a path independent integral defined for deformation theory as a mea
sure ofthe level of deformation at the crack tip. It can readily be shown that these two
approaches lead to identical results as long as a deformation theory of plasticity is
employed. The present approach which accents the use of the plastic intensity factors
is not restricted to use of a deformation theory as has been discussed and may, for this
reason, be somewhat more attractive.

SUMMARY
The initial fracture predictions from classical fracture mechanics for cracked sheets
in tension are quite accurate for applied stresses up to approximately half the yield
stress, even though at half the yield stress the elastic-plastic boundary has been dis
torted considerably from that associated with the small scale yielding solution. For
cracks of shorter length which fracture at higher stress levels, the critical initiation
stress is found to deviate more from the small scale yielding predictions as the strain
hardening capacity is reduced. Very similar conclusions follow from previous work on
the anti-plane shear problem.
The method developed here for combining the knowledge of the dominant singular
solution with the finite element technique to obtain accurate solutions in the neighbor
hood of the crack tip is also applicable to the treatment of problems involving cracks in
finite bodies. It is for unusual geometries that the finite element technique is particu
larly attractive. The application of these techniques to the calculation of elastic stress
intensity factors is straightforward and has been carried out for each of the problems
considered in this study.

REFERENCES

[1] F. A. McClintock and G. R. Irwin, Plastic aspects offracture mechanics. ASTM Spec. Tech. Publ.
No. 381,84(1965).
[2] J. R. Rice, Mathematical analysis in the mechanics of fracture. Fracture Vol. II, pp. 191-311. Academic
Press, New York, (1968).
[3] J. A. Hult and F. A. McClintock, Elastic-plastic stress and strain distribution around sharp notches
under repeated shear. Proc. 9th Int. Congr. Appl. Mech., Brussels. Vol. 8, pp. 51-58 (1956).
[4] H. Neuber, Theory of stress concentration for shear-strain prismatical bodies with arbitrary non-linear
stress strain law. ASME 83-E, 544 (1961).
[5] J. R. Rice, Stresses due to a sharp notch in a work hardening elastic plastic material loaded by long
tudinal shear. J. appl. Mech, 34, 287-298 (1967).
[6] J. R. Rice and G. Rosengren, Plane strain deformation near a crack tip in a power-hardening material.
J. Mech. Phys. Solids 16, (1968).
[7] J. W. Hutchinson, Singular behavior at the end of a tensile crack in a hardening material.J. Mech. Phys,
Solids 16, 13 (1968).
[8] J. W. Hutchinson, Plastic stress and strain fields at a crack tip. J. Mech. Phys, Solids 16,337 (1968).
[9] P. C. Paris and G. C. Sih, Stress analysis of cracks. ASTM Spec. Tech. Publ. No. 381,30 (1965).
[to] W. F. Brown and J. E. Srawley, Plane strain crack toughness testing of high strength metallic materials.
AS1'M Spec. Tech. Publ. No. 410 (1966),

[tl] 1. S. Tuba, private communication.

[12] G. Fix, Higher-order rayleigh-ritz approximations, J. Math. Mech. 18,645-658 (1969).

[13] O. C. Zienkiewicz, The Finite Element Method in Structural and Continuum Mechanics. McGraw
Hill, New York (1967).
[14] B. Budiansky, Extension of rnichell's theorem to problems of plasticity and creep. Q. appl. Math. 16
307-309 (1958).
[15] N. I. Muskhelishvili, Some Basic Problems of the Mathematical Theory ofElasticity. N.V.P. Noord
hoff Groningen, Holland (1963).

(Received 10May 1969)